First of all, the sensors can only pick up signals from the tiny sampling of neurons that immediately surround them. The rest of the brain’s activity remains unreadable, beyond reach. And there’s no preset cipher for the signals they do pick up. Instead, a semblance of understanding between man and machine emerges gradually, through mutual effort.
It starts with Burkhart thinking as hard as he can about performing a certain task, like clasping his hand. (Figuring out exactly what to think about is trickier than it sounds, Burkhart notes. Back when he actually could clasp his hand, he never had to think about it.) As he concentrates, the sensors in his brain send data to a computer for analysis by specially designed machine-learning software. Over time, the software works to pick out an identifiable pattern of activity that corresponds to Burkhart trying to clasp his hand—and then, eventually, to differentiate that from the pattern that corresponds to Burkhart trying to unclasp his hand, or turn it over, or use it to wave hello.
On the software end, there are really three distinct problems to solve, explains Battelle’s Chad Bouton, leader of the Neurobridge project. The first is implanting the chip in the right part of the brain to decode the neural signals. The second is using software to decode those signals. And the third is recoding them and sending them to the sleeve to electrically stimulate the patient’s muscles in the appropriate ways. “We’re basically creating what we believe is going to be a virtual spinal cord,” Bouton says.
Crucially, though, it isn’t just the software that’s learning. Burkhart, for his part, learns by trial and error how to focus his mental activity in a way that the software can understand. That wouldn’t be so hard if the only goal were to reliably perform a single movement, like tapping a single key on a keyboard. But it becomes exponentially more difficult when the goal is to simultaneously modulate the movements of multiple digits and string them together into specific patterns. Add in the fact that he still can’t feel anything in his hands—he can control them, but there’s no sensory feedback mechanism—and Burkhart says he’d be “shocked” if he could pull off something as complex as typing on a keyboard anytime soon.
In the end, the process isn’t so much like one person learning a foreign language. It’s more like two people who speak different languages inventing a new one they can both understand.
For Burkhart, the experience has been as fascinating as it is exhausting. “The guys I’m working with, between the doctors at OSU and the researchers at Battelle—they’re ridiculously smart,” he says. “I have learned a lot just going into those sessions and kind of being a fly on the wall. It’s pretty amazing that I can give them feedback and be a part of it.”
This week, two months after Burkhart’s surgery, Ohio State and Battelle went public with their achievements, billing Burkhart as the first paralyzed man to move his own hand with his thoughts. So far he has mastered five discrete patterns of movement: opening and closing his hand, rotating it up or down, moving it up or down with his wrist, and drumming on a table with his fingers, one after the next. Burkhart goes to the hospital and does these things, and he and the researchers rejoice and marvel at how far he’s come, and videographers record it and send news of the achievement to the world, and then he goes home and is immobile again.
There’s one more catch. As Antonio Regalado poignantly documented in a recent feature story for MIT Technology Review, the chips that researchers implant in patients’ brains aren’t permanent. Rather, the sensors’ reach degrades over time. The result is that Scheuermann, the Pittsburgh woman who made headlines just 18 months ago for feeding herself chocolate with a robotic arm, is already losing some of the dexterity she fought so hard to regain.
Burkhart says he knows the chip will have to come out eventually—probably within five years. But he has faith that experiments like his will lead to further technological advances by then. “It’s already pretty crazy what they’re able to do now. But I wouldn’t go so far as to say we’re living in the future. I know just by looking at history that it’s only going to get better.”
That might seem like a safe bet, but as Regalado’s article makes clear, it’s not a certainty. Not only is the technology invasive, expensive, and impractical, but the potential market—that is, quadriplegics—is too small to excite venture capitalists or corporations with the kind of resources needed to develop and commercialize it.
Battelle’s Bouton emphasizes to me that the Neurobridge research is still in its early stages, with approval so far to work with up to five patients. Provided it goes well, he agrees the eventual goal will be to find commercial applications. There’s some hope, for instance, that the technology could be adapted for rehabilitation of stroke victims, of whom there are about a million a year in the United States alone. “It’s a big unmet need,” Bouton says.
Even better than brain implants would be a system that could read people’s brains from outside their skulls, like an electroencephalography headset. So far, however, external EEG headsets’ performance has not matched their hype, because they can detect only the very crudest signals emanating from the brain. That’s why the highly anticipated ceremonial first kick of the 2014 World Cup—taken by a paraplegic with an EEG headset and a robotic suit—ended up being a bit of a letdown. Yet despite the enormous cost and limited practicality of brain-computer interfaces, Brown University’s John Donoghue pointed out in MIT Technology Review that the first pacemakers were similarly unwieldy.
Burkhart, for his part, is eager to be a spokesman for the technology—which, of course, is exactly what researchers need if they want to attract more funding. Meanwhile, he’s also asking for donations for his own ongoing medical expenses. Of his paralysis, Burkhart says, “I wouldn’t wish that on anybody, but I really can’t complain about the way my life is right now. If nothing else, I can say I was the first person ever to do something, which is an opportunity I never expected to have.”